[Bug gdb/18449] New: GDB crashes on OSX

[Bug gdb/18449] New: GDB crashes on OSX

[andy.somogyi at gmail dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=18449

            Bug ID: 18449
           Summary: GDB crashes on OSX
           Product: gdb
           Version: 7.9
            Status: NEW
          Severity: normal
          Priority: P2
         Component: gdb
          Assignee: unassigned at sourceware dot org
          Reporter: andy.somogyi at gmail dot com
  Target Milestone: ---

When debugging any non-trivial program, such as the python interpreter, or
virtually any other non-trivial program, gdb will hang on OSX with 100% cpu
usage. 

The problem started somewhere after 6.3, as the Apple provided GDB 6.3 works
perfectly. I've verified this problem with GDB 7.9 and 7.7. Also verified this
on 3 different machines, with 3 different OSX versions. 

On OSX 10.7, when using the built in GDB 6.3, no problems are found. However,
when installing GDB 7.9 from homebrew, the problem exists. On two other
machines, with OSX 10.9 and 10.10, has the same problems with GDB 7.7 and 7.9
and 7.9.1  

This is using GDB through eclipse. 

The initial breakpoint in main is always found just fine, and the first 3-4
breakpoints are also found, then after say 3-4 continues, gdb itself will hang
with 100% cpu usage.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug gdb/18449] GDB crashes on OSX

[keiths at redhat dot com]

https://sourceware.org/bugzilla/show_bug.cgi?id=18449

Keith Seitz <keiths at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |keiths at redhat dot com

--- Comment #1 from Keith Seitz <keiths at redhat dot com> ---
(In reply to Andy Somogyi from comment #0)
> When debugging any non-trivial program, such as the python interpreter, or
> virtually any other non-trivial program, gdb will hang on OSX with 100% cpu
> usage. 

If you could attach GDB to that hung (GDB) process and send in a backtrace,
that might prove really helpful. *Very* few of us active developers have
Mac-based systems. [gdb /path/to/gdb -p $pid_of_gdb -ex "thread apply all bt"
should do the trick]

> The problem started somewhere after 6.3, as the Apple provided GDB 6.3 works
> perfectly. I've verified this problem with GDB 7.9 and 7.7. Also verified
> this on 3 different machines, with 3 different OSX versions. 

For the record, 6.3 (Apple's version) is not really FSF GDB. It was heavily
modified (as was GCC) by Apple and its related community. Comparing Apple 6.3
vs FSF GDB therefore only tells me that it is likely some modification that was
never submitted/accepted/committed in upstream FSF GDB.

> On OSX 10.7, when using the built in GDB 6.3, no problems are found.
> The initial breakpoint in main is always found just fine, and the first 3-4
> breakpoints are also found, then after say 3-4 continues, gdb itself will
> hang with 100% cpu usage.

To even begin to help out, we're going to need you to run gdb independently of
Eclipse. That one step alone will help us isolate the problem all the quicker.

Other useful information: What compiler/version are you using? [I'm assuming
Apple's or MacPorts GCC.]

-- 
You are receiving this mail because:
You are on the CC list for the bug.